While the "average" human gene contains about ten times more intron sequence than exon sequence, there are a number of genes that are considerably larger due to the presence of large introns. At the extreme end of this size range are genes of greater than 1,000 kb which encode mRNAs of a few kb. These genes contain introns which often exceed 100 kb in length. Two genes of this type, FHIT and WWOX, are suspected tumor suppressor genes which also contain common chromosomal fragile sites within their large introns. While tumors frequently show expression of aberrant forms of these genes, the mechanism by which these are produced is unclear. Indeed, the expression of such large genes in general is not well understood although they present an extreme challenge to the fidelity of transcription and RNA processing. This application proposes a discontinuous transcription model for the expression of such genes. Two variants of this model are presented which invoke either a mechanism of RNA potymerase skipping, in which large regions of the introns are not transcribed, or trans splicing of multiple transcripts arising from multiple internal promoters. Experiments are proposed which will address the organization of such genes in the interphase nuclei of cells in which they are expressed. First, the time required to synthesize such large genes in dividing cells will be measured. Preliminary data suggests that this time is significantly shorter than expected. Second, the regions of large genes that are actually transcribed will be determined using RT-PCR. Third, the density of RNA polymerases within exonic and intronic regions will be determined using chromatin immunoprecipitation experiments. Finally, the physical distance between exons and introns within interphase nuclei will be measured using fluorescence in situ hybridization as well as two recently developed methods known as RNA-TRAP and chromosome conformation capture.